After finishing Landau's "Statistical Physics" in the afternoon, Lu Qinglong came out of the arcade and found a place to sit down in a nearby park.

(Rereading Chapter 9 of the optics bible... Oh, so that's how it is...)

Random coincidences follow Poisson statistics, and their standard deviation is equal to the square root of the random coincidence count. The statistical significance of the signal-to-noise ratio depends on the precise subtraction of random coincidences.

Last time, Lu Qinglong got stuck here and couldn't move.

He only knew that "accidental coincidences are not fixed numbers." But he couldn't answer any of these questions at the time: what shape the distribution of accidental coincidences actually was, under what conditions its variance converged, or whether to use maximum likelihood or Bayesian methods when extrapolating the true coincidence rate from the observed data.

After reading "Statistical Physics", every question that Lu Qinglong had been stuck on was resolved by a certain theorem.

He slowly closed his eyes, recalling his father's experiment in his mind.

(When the pump light hits the BBO crystal, a pair of entangled photons are born, and they fly towards the two detectors respectively. This is a perfect match.)

(...)

(A large number of independent random events superimposed, the dark count per unit time follows a Poisson distribution).

(It seems that when my father calculated the signal-to-noise ratio, he treated random coincidences as a fixed number that could be substituted into the formula. But random coincidences are not fixed numbers...)

……

Lu Qinglong opened his eyes, his mind like a video recorder, recalling his father's experimental notebook and the data pages inside with incredible detail.

Let's recalculate it using the physical statistics we learned today.

While calculating on his laptop, Lu Qinglong's brain, locked in the optical bible, was not idle either; the remaining threads were also mobilized to assist in the process.

(The remaining pages of Chapter 9 of the Optics Bible are the frequency domain correspondences that conform to technical statistics—the correspondence between cross-correlation functions, baseline height and accidental coincidences, and the frequency domain optimal design that conforms to the window width.)

(I had already encountered this content from an information processing perspective when I was studying "Fundamentals of Random Signal Analysis".)

(These can be connected to the Poisson process framework in statistical mechanics...)

Lu Qinglong immediately opened his schoolbag, picked up a pen, and recalculated on the draft paper.

He used to only know what his father had missed, but now he knows why his father missed it.

My father wasn't ignorant of these theories; it's just that this method of analysis didn't exist at the time.

In 2001, statistical mechanics in the Eastern physics community was mainly focused on solid-state theory and phase transitions, and few people systematically applied it to error analysis of quantum optics experimental data.

To put it simply, my dad was measuring the new world with old standards in an era when the entire methodology of experimentation hadn't yet been updated.

That's why it's so difficult!

That's why even when you see the peak, you can never quite grasp it.

(Dad, you'd better lie down properly now.)

……

night.

In the pitch-black room, the light from the heavy laptop screen was particularly dazzling.

The light from the screen illuminated a delicate, fair-skinned face.

Lu Qinglonggang received a reply notification message as soon as he entered Tianya Community.

I found my dad's old post.

But when he saw the content of the reply post, he almost couldn't hold back.

Reply by: There is a road ahead.

Hello Professor Qing.

Please allow me to address you in this way. From your previous reply, your understanding of the dark counting behavior of the single-photon detector and the timing logic of the coincidence circuit in the SPDC experiment is far beyond the capabilities of an average student or amateur. My supervisor, Professor Zhou, after reviewing your previous replies, also believes that you are very likely a colleague in the field of optics, or an engineer with profound expertise in this area.

I would like to express my sincere gratitude to you in advance. Following your previous suggestions regarding the dark count calibration and constant-ratio timing discriminator, we have confirmed that the original data processing method and time window selection did indeed have significant deviations. The dark count background is no longer completely uninterpretable as before, and the coincidence peaks are much more stable than before.

Although the system signal-to-noise ratio has not yet truly broken through the bottleneck, at least for the first time we have confirmed that this path is not a dead end.

Professor Zhou greatly admires your experimental diagnostic approach. He even said he'd like to invite you for a drink. Of course, he would much rather have the opportunity to meet you in person for a more in-depth discussion. Would that be convenient for you?

With your encouragement, I have decided to continue in this direction. The project has already received support from the 973 Program, all thanks to your help. You may not know, but your few lines of reply were more than just technical advice; they were like adding fuel to a dying fire.

I am currently facing a bottleneck in my next stage of improvement. This likely lies primarily in the quantitative analysis of the remaining low-frequency noise sources—thermal phase drift of the crystal and scattering light from the coating—and the further compression of the overall timing jitter of the probe link. Finding suitable partners in China to discuss these details is extremely difficult.

Therefore, although this request may seem presumptuous, I sincerely hope to invite you to join our project. Whether as a remote mentor or in a manner convenient for you, the mentor and I look forward to receiving your support.

I look forward to your reply. Thank you again for your selfless help.

路远

2001 5 Month 11 Day.

Lu Qinglong was immediately delighted.

(Dad, you're so funny, using your real name online?)

(You'd never guess that Professor Qing is your son!)

As for Qing, that was the online name Lu Qinglong gave himself.

The account used is that of the Fat and Thin Brothers.

Then there are thousands of folded replies below this post.

"There are definitely more talented people online!"

"The ability to pinpoint problems so incisively must belong to a big shot who has returned from overseas."

"His reply is so professional, could he be an old professor who returned from overseas?"

"Could it be a rich guy?"

"I worship the great god."

……

There are hundreds of posts just guessing his identity; the rest are all about worship.

Lu Qinglong thought for a moment, then clicked the reply window with the mouse, and his chubby little hands began typing on the keyboard.

Teacher Lu, I have read your reply.

Don't rush to treat signal-to-noise ratio as a single indicator.

Your most important task now is to separate the noise sources in the system, rather than continuing to merge them into a single "background value" to be subtracted.

First, perform dark counting calibration. Block the pump light and record the count rates of detectors A and B under different high-voltage biases and temperatures for at least several hours. Do not just take the average value; also record the fluctuation range.

Second, re-examine the coincidence window. If the window is too wide, random coincidences will increase; if the window is too narrow, true coincidences will be excluded by detector timing jitter. Do not start with an extremely narrow window; instead, begin with a wider window and gradually narrow it, recording the changes in the true coincidence count and the background of random coincidences.

Third, do not continue to use the original Gaussian distribution to directly hit the crystal with the pump beam. Try shaping the Gaussian beam into an approximately flat-topped beam to reduce the inhomogeneity of the effective interaction region within the crystal. This step may not immediately improve the signal-to-noise ratio, but it can help determine whether the current bottleneck stems from spatial mode mismatch.

First, do these three sets of data.

If the three sets of data correspond to each other, it means that the route is feasible; if they do not correspond, then we can talk about changing the hardware.

Qing

Okay, press Enter!

Lu Qinglong turned off the monitor and crawled back into bed.

Tap the screen to use advanced tools Tip: You can use left and right keyboard keys to browse between chapters.

You'll Also Like